Dr. Daniel Barulli explores how agentic AI systems are evolving—from tool use to metacognition—and what that means for research, work, and safety.